5V Voltage Regulator

[TP380Z]

Alternators do use a type switch-mode regulation where the Alternator Voltage Regulator switches the voltage to the field winding (alternator rotor) fully on and then turns it off. Unlike the SMPS Buck Regulator the OP is likely to buy from EBay, the AVR switches on/off 30 to 150 times per second, which causes no detectable RFI at 0.5 to 1.5MHz, partly because the current through the alternator rotor is continuous due to the huge inductance of the rotor winding. The spectrum of the AVRs switching rate is such that you would be receiving the several-thousands harmonic at 1Mhz, which has no detectable energy...

Contrast that to an SMPS that switches at 250KHz. It is really hard to suppress its fourth harmonic to prevent it from clobbering 1MHz..., tin boxes notwithstanding....

I was not aware of of the SMPS Buck Regulator. I was originally going to follow the example set out the Allpar Get Engauged article. However, I discovered that the Texas Instruments #UA7805C (DK 296-1974-5-ND) is out-of-stock at Digi-Key which I was going to substitute with DK 497-15681-5-ND. The Allpar article was unclear about the capacitor so, while researching this further, I came across the Dimenision Engineering DE-SW050 switching regulator, which has the same TO-220 3-pin configuration.

Alternators can and do generate some RFI, but it it is wide-band hash due to brush arcing as the brushes ride on the copper slip rings in the rotor. This can be suppressed by mounting a shunt capacitor (with very short leads) right on the alternator. These are effective in suppressing radiated RFI in the 100kHz to several MHz range. This capacitor has nothing to do with filtering alternator ripple, which is an entirely different kettle of fish...

I have a lot of experience installing LORAN and ADF VLF receivers in boats and aircraft. LORAN utilized 100kHz while ADF uses 190kHz to 1.5MHz, and if the alternator filter was missing, reception range would be diminished. Listening to an AM car radio, if the alternator filter is missing, the alternator hash can be heard as a bacon frying sound, especially while tuning weaker AM radio stations.

I spliced-in the entire underhood wiring harness from a 72 Scamp. I have an external capacitor at the ignition coil but nothing at the alternator. I think there may be more noise in my AM radio when the hood up but it quiets down when the hood is closed. It probably wouldn't hurt to add a RC-11 capacitor at the alternator.

Just one word of caution. The sender voltage in a car is safety critical. As you probably know, the sender in the petrol (gas) tank is a wire wound potentiometer with a law tailored to the geometry of the tank. The danger is that the wiper of the potentiometer could cause a spark with obvious consequences. There are regulations, I cant remember them now, about the nature and amplitude of any voltage in a flammable environment, so this is an area to be aware of.

Last year, due to rust, I replaced the gas tank and sender so both are now brand new. I have a Spectra Premium FG69B sender.

 

[MikeMl]

The spectrum from a square wave depends on the speed of the rise time and fall time of the edges not on the PRF. ..

spec

No. Here are the spectra (in the vicinity of 1000KHz, in the center of the AM Broadcast Band) of two waveforms: one is 1Vpp, 100Hz, 30% duty cycle, 50ns rise and fall times (like the AVR), the other is 1Vpp, 250kHz, 30% duty cycle, 50ns rise and fall times (like the typical Buck converter).



Note that the if you had the AM receiver tuned to ~1000KHz, the forth harmonic of the 250KHz switching rate is only about --22db, while the ten-thousandth harmonic of 100Hz is about -89db, for a -67db difference.

The math and practical experience says that it is much easier to live with the AVR than it would be to live the Buck converter, doing nothing but installing the RFI bypass cap on the alternator.

Putting it another way, by the time you do enough shielding and filtering on the Buck converter for an AM receiver to be able to coexist with it, you will have spent 10X time and money on the project as compared to just going with the LM7805...

 

[spec]

spec

No. Here are the spectra (in the vicinity of 1000KHz, in the center of the AM Broadcast Band) of two waveforms: one is 1Vpp, 100Hz, 30% duty cycle, 50ns rise and fall times (like the AVR), the other is 1Vpp, 250kHz, 30% duty cycle, 50ns rise and fall times (like the typical Buck converter).

View attachment 98967 View attachment 98968

Note that the if you had the AM receiver tuned to ~1000KHz, the forth harmonic of the 250KHz switching rate is only about --22db, while the ten-thousandth harmonic of 100Hz is about -89db, for a -67db difference.

Hy Mike,

I must say that your graphs are fairly conclusive, but something worries me. It is my understanding, like I said, that it is the edge that counts. Without going into the maths, if you put a square wave signal, of any PRF, with a fixed rise time into a scope with a fixed bandwidth the edge will be equally degraded indicating that the frequency spectrum, in absolute value rather than relative to the fundamental frequency of the square wave, has the same hf content.

I must admit tough, that I am no expert in this area.

It will be interesting if the OP does fit the switching regulator. Then we will have the proof of the pudding one way or another.

By the way, I am not suggesting that the switching regulator be fitted in the dial on the dash. Instead I would bring out two wires for the 5V supply and mount the switching regulator, with its filtering, under the bonnet (hood).

The math and practical experience says that it is much easier to live with the AVR than it would be to live the Buck converter, doing nothing but installing the RFI bypass cap on the alternator.

Putting it another way, by the time you do enough shielding and filtering on the Buck converter for an AM receiver to be able to coexist with it, you will have spent 10X time and money on the project as compared to just going with the LM7805...

As I have said before, I fully agree that the linear regulator is the best way to go but the OP has stated that he does not want to do that.

spec

 

[spec]

Last year, due to rust, I replaced the gas tank and sender so both are now brand new. I have a Spectra Premium FG69B sender.

Hy TP,

Perhaps I didn't explain what I meant clearly. It is the voltage that goes to the sender in the gas tank that is safety critical, rather than the sender itself. Suppose for the sake of argument, that the regulator failed in such a way that it put out the full 12V line voltage, which might be around 14.7V when the alternator is running. That voltage might be sufficient to ignite the fuel air mixture in an almost empty tank.

spec

 

[spec]

I spliced-in the entire underhood wiring harness from a 72 Scamp. I have an external capacitor at the ignition coil but nothing at the alternator. I think there may be more noise in my AM radio when the hood up but it quiets down when the hood is closed. It probably wouldn't hurt to add a RC-11 capacitor at the alternator.

If you pick up more hash with the hood open it tends to indicate radiation from the ignition circuit rather than the alternator. Do you have resistor leads and even resistor plugs fitted. Does the Barracuda have mechanical points or electronic ignition?

spec

 

[TP380Z]

Hy TP,

Perhaps I didn't explain what I meant clearly. It is the voltage that goes to the sender in the gas tank that is safety critical, rather than the sender itself. Suppose for the sake of argument, that the regulator failed in such a way that it put out the full 12V line voltage, which might be around 14.7V when the alternator is running. That voltage might be sufficient to ignite the fuel air mixture in an almost empty tank.spec

Both my gasoline and temperature gauges are supplied by the 5V mechanical voltage limiter. Each gauge has 20 Ω which then leads to the senders. I'm not sure of the temperature sender but the fuel sender is 80-10 Ω, and I believe full scale for both senders is 10 Ω. If the mechanical instrument voltage limiter fails and supplies the full 14 volts to the gauges, the gauges would fry first.

In my case, I suspect that my voltage limiter is on its way out because both of my gauges seem to read a bit low sometimes.

I'm not against using a 7805 but the 497-15681-5-ND costs $US$0.93 while the DE-SW050 costs US$15.00, which are still cheaper than an external mechanical voltage limiter from a newer vehicle.

As for my ignition system, I am using a mopar electronic system with oem-style resistor wires and plugs.

 

[spec]

Hy again Mike,

I have just done a bit of revision and, yes you are right. With a perfect square wave as an example, the amplitude of the even harmonics is zero and the formula for odd harmonics is:

where:

• An = the amplitude of the nth harmonic
• pi = the constant, 3.14159...
• n = the harmonic number (only odd)

spec

 

[MikeMl]

The reason I used 30% duty cycle is to avoid the even harmonics issue

The energy in the nth harmonic is inversely proportional to n, so it stands to reason that the ten-thousandth harmonic would be way down compared to the fourth harmonic, the rise/fall times being equal...

 

[TP380Z]

I asked DE about RFI from their DW-SW050 and their response was:

AM should be outside the range of this regulator's frequency, so you should be fine. But if you want to add some noise filtering, page 5 of the datasheet shows how this is done:
DE-SW0XX Data Sheet

According to the datasheet, the regulator has a switching frequency between 230-290 kHz, with the typical being 270 kHz.

I see that DE recommends putting the 470 μF capacitor on the output side of the regulator.

 

[MikeMl]

...According to the datasheet, the regulator has a switching frequency between 230-290 kHz, with the typical being 270 kHz.

I see that DE recommends putting the 470 μF capacitor on the output side of the regulator.

Go ahead, then tune your AM radio from 540kHz to 1650kHz, and tell us how many loud buzzes you hear...

 

[bountyhunter]

"According to the datasheet, the regulator has a switching frequency between 230-290 kHz, with the typical being 270 kHz.

I see that DE recommends putting the 470 μF capacitor on the output side of the regulator."

With no information on the capacitor? If they are actually trying to bypass RFI and switching ripple, that cap would have to be one that can handle the 300 kHz frequency. Most aluminum electrolytics would be useless, their self resonant freq is lower than 300 kHz. Even Tantalums tend to be going away at about 300 - 400 k ballpark. Probably need to use "switcher" Tantalums bypassed by large ceramics to do any good at actually killing the RFI.

That data sheet is a joke.

 

[shokjok]

I would use a TO-3 cased 78K05 voltage regulator to make up for the resistance loss to the module. A good heatsink for this device would be on the firewall near the tranny shift linkage, away from the exhaust pipes, unaffected by moisture. From the schematics I've seen, there is no direct connection to a switched 12V source as these devices are fed a reduced voltage from the ignition computer.